The 40Ar/39Ar step-heating analyses of three hornblende samples from the Border Ranges ultramafic-mafic assemblage yield plateaus with slightly discordant dates of 180.2¿0.7, 177.5¿0.7, and 181.1¿1.7 Ma. A plot of the plateau temperature data for all three hornblende samples on an 36Ar/40Ar versus 39Ar/40Ar diagram reveals the presence of a non-atmospheric ''initial'' argon component with an apparent 40A6/36Ar of 379¿16 and yields a date of 175.7¿1.0 Ma. A potassium-poor hastingsite from the same complex yields a complicated spectra with dates as high as ~900 Ma. These older dates reflect contamination by argon with an isotopic composition similar to that found for the initial argon in the hornblende and do not record a Palozoic component in the age of the Border Ranges ultramafic-mafic complex. Hornblende from an Eocene tonalite sill in the nearby Chugach Metamorphic Complex also contains an initial argon component (40Ar/36Ar=314¿4). The composition of this initial argon is presumably dictated by the argon isotopic composition of the host rock fluid at the time of amphibole/fluid equilibration. Approximately 2¿10-7 cm3/g of this initial argon is sufficient to perturb the 40Ar*/39ArK ''model'' plateau dates. Although 40Ar/39Ar stepwise degassing at low temperatures removes an additional ''atmospheric'' argon component, the initial argon in these samples is released at a rate that closely coincides with the degassing rate of radiogenic argon from the amphibole, and only through application of the 36Ar/40Ar versus 39Ar/40Ar diagram can this nonatmospheric argon component be detected. For the Border Ranges ultramafic-mafic and Chugach Metamorphic complexes, the range in K-Ar amphibole dates could be interpreted as recording differential uplift and an extended duration of intrusive activity, respectively. The results suggest, however, that nonatmospheric initial argon contributes significantly to the observed age variations. ¿ American Geophysical Union 1989